Because of the randomness

intermittence and uncertainty

the integration of large-scale wind farms in a power system will exert some influence on the distribution of the power flow. Like most studies in the past

the target is always modeling the dependence structure between different wind farms geographically. In view of the situation in Yunnan power grid where a large wind farm power grid is often composed of a number of wind turbine groups scattered on different hilltops

it is no way to simply use one equivalent wind turbine to model the whole wind farm any more. Based on the theory that the Copula function is able to describe different tail characteristics

the mixed Copula function

often abbreviated to M-Copula

is proposed to model the wind farm’s output.As one of the generator nodes in the western Yunnan power grid

a certain wind farm is chosen for empirical analysis. The result shows that the M-Copula function can play almost the same role in wind farm power modeling as the actual turbines’ output on the platform of PSD-BPA

which single Copulas cannot.